Kaufman ZA, Spencer TS, Ross DA. Effectiveness of sport-based HIV prevention interventions: a systematic review of the evidence. AIDS Behav. 2012 Oct 25. [Epub ahead of print]
To evaluate the effectiveness of sports-based HIV prevention (SBHP) interventions and to identify gaps and limitations of existing research.
Randomized controlled trials (RCTs); non-randomized interventions; cross-sectional or case-control studies with intervention and control groups.
Not stated, but apparently adolescents or adults in any setting.
Interventions that explicitly used sports themes, activities, metaphors, and/or role models in an effort to reduce HIV transmission. This included educational interventions as well as interventions aiming to increase uptake of health services that contribute to reducing HIV transmission and/or acquisition, e.g. medical male circumcision (MMC), HIV counseling and testing (HCT), sexually transmitted infection (STI) treatment, or antiretroviral therapy (ART). Studies that did not assess effectiveness quantitatively were excluded.
HIV knowledge; HIV stigma; self-efficacy; reported communication (i.e. ability to communicate with others about HIV); self-reported sexual risk behavior, uptake of HIV services; incidence or prevalence of HIV, STIs, or pregnancy.
EMBASE, "Global Health," MEDLINE and PsycINFO were searched in August 2011, using a range of relevant sports terms, study design terms and terms relevant to HIV, STIs and pregnancy. National Library of Medicine (NLM) Medical Subject Heading (MeSH) terms were apparently not used in MEDLINE. Researchers in the field and staff at key organizations were contacted and asked for relevant published or unpublished study data. It is not stated whether articles published in all languages were eligible for inclusion.
The authors identified 924 records. They excluded 869 duplicate or irrelevant records, and then reviewed the abstracts of 83 studies. They examined the full texts of 29 articles. Twenty-one studies were included in the review. The authors do not report whether the screening process was performed by two or more reviewers working independently, nor do they report details of their data extraction process.
No RCTs of SBHP interventions were identified. Seven of the included studies were quasi-experimental (i.e. non-randomized, prospective studies with intervention and comparison groups); four were cross-sectional, and ten utilized a pre/post (or time-series comparison) design. Sixteen of the studies were conducted in sub-Saharan Africa, two in the Caribbean region, and three in the United States. Most study participants were between 12 and 16 years old; only one study included participants age >30 years. Most of these studies assessed youth-targeted interventions delivered either with sports teams or school classes through curricula that used sports themes, activities, and metaphors. All identified studies were published between 2006 and 2011.
Eighteen studies assessed HIV knowledge, 14 assessed reported attitudes toward HIV, five assessed reported communication about HIV, six assessed reported sexual risk behavior and four assessed uptake of HCT. No studies assessed biological outcomes, and no studies assessed uptake of other HIV services.
Study findings were analyzed within outcome categories to determine the evidence of intervention effectiveness in improving knowledge, stigma, self-efficacy reported communication, reported sexual behavior, service uptake, and biological outcomes. Random-effects meta-analyses were performed for overall effects on these outcomes. Two levels of sensitivity analysis were carried out, the first by excluding studies classified as "poor quality" and the second by excluding unpublished studies.
Results by outcome: HIV knowledge: Of the 18 studies assessing HIV-related knowledge, 15 reported an overall positive effect, two found no effect and one found a negative effect. Of the 15 studies finding a positive effect, 10 had very strong evidence of effectiveness (p<0.01). Meta-analysis for this outcome found strong evidence of positive effects across studies (relative risk [RR] 1.25, 95 % confidence interval [CI] 1.16 to 1.34). Stigma: Eight of 11 studies assessing stigma reported a positive effect. Meta-analysis for this outcome found strong overall evidence of effectiveness across studies (RR 1.21, 95 % CI 1.09 to 1.32). Self-efficacy: Three of five studies reported increases in self-efficacy. Meta-analysis for this outcome found strong evidence of effectiveness across studies (RR 1.22, 95 % CI 1.02 to 1.41). Attitudes: Five of eight studies reported a positive effect on other HIV-related attitudes and life skills. Meta-analysis for this outcome was apparently not performed. Communication: Of the five studies assessing reported communication, four studies found a positive effect on reported HIV-related communication and one multi-country study found no effect. Meta-analysis for this outcome across studies found strong evidence of effectiveness (RR 1.24, 95 % CI 1.06 to 1.41). Sexual risk behavior: Of the six studies that assessed an effect on reported sexual behavior, five reported positive effects on at least one behavior, and one study found no effect. Meta-analysis for recent condom use across studies found strong evidence of an effect (RR 1.29, 95 % CI 1.00 to 1.59). No studies found evidence of a negative effect on reported behaviors. Service uptake: Three studies measured intervention effectiveness in increasing reported uptake of HCT. No evidence of an overall effect on HCT uptake was observed in the meta-analysis (RR 1.81, 95 % CI 0.20 to 3.42), though two of the three studies found very strong evidence of effectiveness.
Summary of meta-analyses, including sensitivity analyses, across outcomes. (From the article)
The authors conclude that that well-designed and implemented SBHP interventions can significantly reduce stigma and significantly increase HIV-related knowledge, self-efficacy, reported communication and condom use. They find no evidence that these interventions can reduce HIV, STI or unintended pregnancy rates, though they also find no evidence to suggest harms or negative effects in these interventions.
Study quality was assessed using a modified version of the Newcastle-Ottawa Quality Assessment scale (NOS).1 The adapted scale included all study design elements from the NOS (related to selection, comparability and outcome measurement) with additional elements accounting for random allocation, use and reporting of appropriate sample size calculations, sufficient study size, and appropriate statistical analysis. Two studies were classified as "Good-quality", 11 as "OK-quality," and eight as "Poor-quality." None of the included studies were classified as having "Very Good quality." The mean study quality score was 5.1 (standard deviation [SD] 3.1) points out of 20 possible points. The most common limitations identified were lack of objective outcome measures (all studies relied on self-reported outcomes), lack of randomization in both sampling and group allocation, and lack of extended follow-up. Only five studies adjusted analyses for confounders, only three reported that they had used sample size calculations to determine the study size, and only one followed participants for more than six months.
This quality of this systematic review is high, although it does not provide a review protocol and the screening and data extraction process was not well described. Based on the overall quality of the review, however, it seems likely that many of these deficiencies were in the original reports, and that the review itself was conducted appropriately. The review meets nearly all the other criteria of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist.2
The majority of studies identified in this review only assessed knowledge and stigma, and none assessed biological outcomes. This is a major limitation in assessing SBHP effectiveness. Since knowledge and reported attitudes are quite distal factors in influencing HIV risk, it is necessary to use caution in interpreting the true efficacy and health impact (i.e., reduction in new HIV infections) of these interventions.
Researchers and public health practitioners designing HIV prevention interventions may wish to include sports-based components, or to model them on existing SBHP interventions.
- Wells G, Shea B, O'Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. [accessed 27 October 2012]
- Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009) Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097